Israeli Military Used Lavender AI to Pick Gaza Targets, With 10% Misidentified
Updated
Updated · The Guardian · Oct 2
Israeli Military Used Lavender AI to Pick Gaza Targets, With 10% Misidentified
3 articles · Updated · The Guardian · Oct 2
Summary
Lavender flagged suspected Hamas operatives from phone-data patterns after the Oct. 7, 2023 attack, and Israeli officers often approved those AI-generated targets in about 20 seconds, according to the report.
Where’s Daddy? then tracked suspects to their homes, making family members vulnerable because strikes were aimed where targets slept rather than at active military sites.
Israel’s military has acknowledged roughly 10% of people marked for assassination were not Hamas military-wing members, while the system still fed a rapid target pipeline.
The report says commanders accepted an expected toll of 20 civilian deaths for a suspected low-level fighter, a ratio it argues would breach proportionality under international humanitarian law.
The Gaza case is presented as a warning for Geneva talks on autonomous weapons, where efforts to preserve meaningful human review of AI-selected targets remain contested.
If AI hallucinations can nearly trigger global conflicts, are militaries moving too fast to automate the kill chain?
With treaties lagging behind tech, who holds ultimate moral responsibility when an autonomous drone makes a deadly mistake?
Algorithmic Warfare and Civilian Harm: The Global Legal Crisis Over AI-Driven Targeting in Gaza and Beyond (2024–2026)
Overview
The report reveals how Israel’s use of the Lavender AI system in Gaza led to high civilian casualties, as human oversight was reduced and large numbers of targets were approved quickly. Human rights groups and the ICRC warned that faulty data and automation bias increased the risk of mistakes, while legal experts highlighted a growing accountability gap as operators blamed algorithmic errors. The resulting international backlash included ICC arrest warrants for Israeli leaders and diplomatic actions like Spain canceling arms contracts. Meanwhile, similar AI-driven targeting in the U.S. military led to tragic errors, showing how powerful nations resist global efforts to regulate autonomous weapons and protect their military AI capabilities.